Engineering Subtilisin for Peptide Coupling - American Chemical

Jun 15, 1994 - Engineering Subtilisin for Peptide Coupling: Studies on the .... 0. 0. 50. 100. 150. Time (hours). Figwe 1. Stability of subtilisin BPN...
0 downloads 0 Views 2MB Size
J. Am. Chem. SOC.1994,116, 6521-6530

6521

Engineering Subtilisin for Peptide Coupling: Studies on the Effects of Counterions and Site-Specific Modifications on the Stability and Specificity of the Enzyme Pamela Sears, Matthias Schuster, Peng Wang, Krista Witte, and Chi-Huey Wong’ Contribution from the Department of Chemistry, The Scripps Research Institute, 10666 North Torrey Pines Road, La Jolla, California 92037 Received January 28, 1994“

Abstract: Several variants derived from the thermostable subtilisin 8397 were made in order to create an enzyme that is more stable toward organic solvents or has a broader specificity for the PI’ residue in amidation or is more effective for peptide segment ligation in aqueous solution. To improve the stability in organic solvents, one of three surface charges was removed each time from 8397 to create the variants: Lys43 Asn (K43N), Lys256 Tyr (K256Y), and Asp181 Asn (D181N). Although the stabilities of these variants in high concentrations of hydrophilic organic solvents were higher than that of the wild-type enzyme, the D181N variant was less stable than the 8397 variant. It appears that removal of isolated surface charges does not necessarily improve the enzyme stability in polar organic solvents. A dramatic change of the enzyme stability in dimethylformamide (DMF) was, however, observed in the presence of different counterions. Subtilisin BPN’ lyophilized from Tris-HC1 buffer (50 mM, pH 8.4) and suspended in DMF (solid partially soluble), for example, was completely inactivated in 30 min at 25 OC, while the enzyme still retained about 70% of the original activity in a week if lyophilized from sodium phosphate buffer (50 mM, pH 8.4) (solid completely insoluble in DMF). In general, the enzyme lyophilized from organic buffers deactivates in DMF much faster than that from inorganic buffers. A similar counterion effect was observed with other variants. These studies suggest that subtilisins are very unstable when exposed directly to DMF; the stability is, however, markedly improved if the enzyme is protected by water or salts from contact with the solvent. To use subtilisins and variants in transesterification or aminolysis in organic solvents, water (3-30%) is usually present in order to have significant reactivity, and for transesterifications, it was found that a good rate and yield could be achieved in ethanol containing 30% water. For use in peptide segment ligation in aqueous solution, the active-site serine of subtilisin 8397/C206Q was converted chemically to cysteine, forming thiosubtilisin 8397/C206Q, and the amino1ysis:hydrolysis ratio was found to be several orders of magnitude higher than that for subtilisin BPN’ and comparable to that for thiosubtilisin BPN’. The 8397 variant was also modified at the SI’ site via M222A/Y217W mutations to broaden the PI’ specificity.

-

-

Introduction A great deal of interest has been focused in recent years on the use of proteases in peptide synthesis.’ While solid-phase2 and solution-phase3peptide syntheses have reached a high degree of sophistication, they still require extensive protection and deprotection steps and suffer from racemization and low solubility of many derivatized amino acids and peptides in the solvents used; the methods are therefore not generally applicable to the synthesis of large peptides or proteins or their glycoconjugates. Although enzymatic peptide coupling has been considered to be potentially useful and to have many advantages, including greater stereoand regioselectivity, reduced need for protecting groups, and mild reaction conditions, it is still not commonly practiced in the laboratory. Part of the reason is that proteases also catalyze the hydrolysis of both the substrates and the peptide products, and the synthesis of peptide esters for the kinetically controlled peptide ligation in aqueous solution is not trivial. Moreover, many proteases are unstable toward polar organic cosolvents, which often must be used for small protected substrates both to decrease the rate of Abstract published in Advance ACS Abstracts, June 15, 1994. (1) Kullmann, W. Enzymoticpeptide synthesis; CRC Press: Boca Raton, FL, 1987. Schellenberger, V.; Jakubke, H.-D. Angew. Chem., Int. Ed. Engl. 1991, 30, 1437. Wong, C.-H.; Wang, K.-T. Experientia 1991, 47, 1123. Chaiken, I. M. Crit. Rev. Biochem. 1981, 21, 255. (2) Merrifield, R. B. Angew. Chem., Int. Ed. Engl. 1985,24,799. Kent, S . In Innovation and Perspectives in Solid Phase Synthesis; Epton, R., Ed.; SPPC Ltd.: Birmingham, U.K.,1992. (3) Bodanszky, M. Principles ofpeptide synthesis; Spring-Verlag: Berlin, 1984.

-

hydrolysis and increase the solubility of the substrates. Our goal is to create stable enzymes that overcome these problems and can be used in aqueous solution with or without organic cosolvents for kinetically controlled (i.e. aminolysis of esters) peptide synthesis. Starting with the thermostable subtilisin 8397,4,5we have attempted to createvariants that are stable towardorganic solvents by removing surface charges, an approach that has shown some success in the past with a-lytic protease? The residues chosen for site-directed mutagenesis were those which had no interaction with other nearby charged residues (or at least were not in locally neutral regions), as determined by analysis of the crystal structure, and were not conserved in other subtilisin variants. The residues thus chosen were changed to the amino acids found most commonly in the equivalent positions in the other subtilisinvariants surveyed. These variants were studied for their stability and kinetics in buffer/organic cosolvent mixtures. In addition, the (4) For early work on the preparation of thermostable subtilisin 8350, see: Pantoliano, M.; Whitlow, M.; Wood, J. F.; Dodd, S. W.;Hardman, K. D.; Rollence, M. L.; Bryan, P.N. Biochemistry 1989, 28, 7205. This enzyme is modified from subtilisin BPN’ through the following changes: MetSOPhe, Asn76Asp,Gly169Ala,Asn218Ser,Gln206Cys,andTyr217Lys.Theenzyme is stable in high concentrations of DMFand useful for synthesis: Wong, C.H.; Chen,S.-T.;Hennen,W.J.;Bibbs,J.A.;Wang,Y.-F.;Liu,J.L.-C.;Pantoliano, M.W.; Whitlow, M.; Bryan, P. N. J. Am. Chem. Soc. 1990, 112, 945. ( 5 ) For the preparation and application of subtilisin 8397, see: Zhong, Z.; Liu, J. L.-C.; Dinterman,L.M.; Finkelman,M. A. J.; Mueller, W. T.; Rollence, M. L.; Whitlow, M.; Wong, C.-H. J. Am. Chem. Soc. 1991,213,683. The situ of modifications are indicated in Figure 9. (6) Martinez, P.; Arnold. F. J . Am. Chem. SOC.1991, 223, 6336. Chen, K.; Arnold, F. H. Proc. Nail. Acad. Sci. U.S.A. 1993, 90, 5618.

0002-7863/94/1516-6521$04.50/00 1994 American Chemical Society

6522 J. Am. Chem. SOC.,Vol. 116, No. 15, 1994

Sears et al.

Table 1. Kinetic Parameters of Subtilisin Mutants in 100 mM Tris, pH 8.0, T = 25 OC

wt

8397

K43N

K256Y

50 m M Sodium Phosphate

D181N

sAAPFpNAa kat (SI) Km (mM) kcat/Km

68 0.42 160

100 0.18 560

60 0.19 315

90

110

0.11 47 0

0.2 550

(mM-l s-I)

sAAPFSBzb

kcat (s-')

Km (mM) kcat/Km

2.8 x 103 1.3 x 103 1.2 x 103 1.3 x 103 1.2 x 103 0.62 0.34 0.27 0.24 0.27 4.5 x 103 3.8 x 103 4.4 x 103 5.4 x 103 4.4 x 103

0.2 O - 1

(mM-l s-I)

0.

a Succinyl-Ala-Ala-Pro-Phe p-nitroanilide. Succinyl-Ala-Ala-ProPhe thiobenzyl ester.

effect of counterions on the enzyme stability in polar organic solvents such as dimethylformamide(DMF) was also investigated. In order to increase the ratio of peptide bond formation to substrate hydrolysis (theaminolysis to hydrolysis ratio) in aqueous solution, thios~btilisin~-~ has been used. The enzyme, however, only accepts activated peptide esters as substrates. When peptide methyl esters are used as substrates, the reaction must be conducted at higher temperature. The thioenzyme is, however, very unstable at >50 "C (t1p < 1 h). We therefore converted the active-site serine of stable subtilisin variants to cysteindOto be used at high temperatures in peptide coupling with peptide methyl esters. We have also prepared two variants from 8397 that exhibit a broader PI' specificity in peptide coupling.

50

0

100

150

Time (hours) Figwe 1. Stability of subtilisin BPN' lyophilized from different buffers atpH8.4andsuspendedinanhydrousDMF(T=25 "C,withoutstirring). 0.51

@

stirred

Results

StableSubtilisinVariants in Polar OrganicSolvents. Following chromatography on DEAE cellulose, the purity of the mutant subtilisin preparations was greater than 90%,as assayed by SDSPAGE. Active-site titrations with cinnamoylimidazole demonstrated that approximately 60-704 of the protein was active. The kinetic constants of the subtilisin variants after charge removal were measured with the hydrolysis of the amide and ester substrates sAAPFpNA and sAAPFSBz. The results are essentially the same as for subtilisin 83975 (Table I), indicating no gross conformational changes, except that small changes for k,,, and K, are observed. We first attempted to measure the stabilities of the subtilisin mutants in anhydrous organic solvents, since the 8397 variant has been reported to be more stable than the wild-type enzyme under high concentrations of DMF.S Aliquots of enzyme were taken from the suspension over a period of time and assayed in aqueous solution. The initial velocity of the enzyme activity was used as a measure of the stability. We found, however, that the stability of the enzymes in anhydrous solvents is extremely dependent on the amount and type of salts associated with the enzyme. The enzyme prepared via lyophilization from a sodium phosphate buffer and suspended in anhydrous DMF is markedly more stable than that prepared from a Tris buffer under the same conditions (Le. same pH and buffer concentration). The wildtype enzyme with Tris was completely inactivated in 30 min while the enzyme associated with phosphate retained 70% of the activity after 1week when unstirred (Figure 1). Furthermore, thestability is even dependent on the stirring speed (Figure 2). A similar situation was observed in both 99% DMFand 90%DMF (balance (7) For useof thiosubtilisinin peptidesynthesis,see: Nakatsuka, T.; Sasaki, T.; Kaiser, E. T. J . Am. Chem. Soc. 1987, 109, 3808. (8) The ratio of aminolysis to hydrolyis is increased by a factor of -7000: (a) Wu, 2.-P.; Hilvert, D. J . Am. Chem. Soc. 1989, I l l , 4513. (b) Zhong, Z . ; Wong, C.-H. Biomed. Biochim. Acta 1991,10/11, S9. (9) Further improvement of aminolysis was observed with a double mutant (Ser221- Cys. Pro225 +Ala): Abrahmsbn, L.; Tom, J.; Burnier, J.; Butcher, K. A.; Kossiakoff, A.; Wells, J. A. Biochemistry 1991, 30, 4151. (10) For application of thermostable thiosubtilisin variants to glycopeptide synthesis, see: Wong, C.-H.; Schuster, M.; Wang, P.; Sears, P. J. Am. Chem. SOC.1993, 115, 5893.

" I

0

250

500

750

lo00

Time (hr.)

Figure 2. Effect of stirring on the stability of subtilisin 8397 in 100% DMF (enzyme lyophilized from 10 mM sodium phosphate/0.025 mM CaC12, pH 7.0; assay conducted at T = 25 "C). Table 2. Stability of Subtilisin BPN' in 90% and 99% DMF (Balance Water) as a Function of the Counterion Type and

Concentration 99% DMF

90% DMF t(50%)b(min) (A) Subtilisin Lyophilized from 10 mM of Various Buffers + t(50%)*

HEPES TIiS

0.025 mM CaC12'